Crack image recognition and modeling method

A modeling method and image recognition technology, applied in the field of image recognition, can solve the problems of difficult to extract accurate coordinates, difficult to mark small cracks, and low recognition accuracy, so as to improve the recognition accuracy, save labor costs, and improve robustness. high effect

Active Publication Date: 2019-12-31
NANJING UNIV OF SCI & TECH
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the unique narrow width of cracks and the difficulty of marking small cracks, the existing neural network identification methods are poor in crack identification, and the accuracy of identification is low.
At present, the neural network is more suitable for qualitative image targets, and it is difficult to extract accurate coordinates
The identification of the boundary area is difficult to meet the needs of the subsequent 3D modeling

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Crack image recognition and modeling method
  • Crack image recognition and modeling method
  • Crack image recognition and modeling method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026] The implementation of the present invention will be described in detail below in conjunction with specific embodiments.

[0027] A crack image recognition and modeling method, comprising the following steps:

[0028] Step 1. Rough recognition of crack images based on deep learning Mask-RCNN network and rough extraction of crack images

[0029] Step 1.1: Prepare crack pictures for training and crack pictures to be detected for training the network, the process is as follows.

[0030] Prepare a number of crack pictures with the same pixel size, and prepare pictures of multiple angles for each crack. Divide these crack pictures into two parts, one part is used as a training crack picture, and the other part is used as a crack picture to be detected, which can effectively improve the recognition accuracy. Rate.

[0031] Step 1.2: Use labelme to label the crack pictures for training, the process is as follows:

[0032] First, import the training crack picture into labelme...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a crack image identification and modeling method. The method comprises the following steps: 1, carrying out coarse identification on a crack image based on a deep learning Mask-RCNN network and obtaining a coarse extraction crack image; 2, performing OpenCV image processing on the roughly extracted crack image to obtain pixel coordinates of inflection points and abrupt change points of the cracks on the two-dimensional image; 3, performing image feature matching by using an SIFT algorithm to obtain coordinates of homonymous points; and 4, constructing a three-dimensional model of the crack. According to the invention, the advantages of the deep learning image recognition technology and the OpenCV image processing technology are combined; the image recognition methodfrom coarse recognition to accurate recognition is formed, the recognition accuracy is effectively improved, crack three-dimensional modeling is finally generated by means of extraction of picture coordinates of different angles at the same crack position, the problem that cracks are difficult to recognize and model due to small sizes is solved, and the method is high in feasibility and robustness.

Description

technical field [0001] The invention belongs to the technical field of image recognition, and in particular relates to a crack image recognition and modeling method. Background technique [0002] Image recognition and modeling is to realize the extraction and three-dimensional reconstruction of image targets through a series of processing processes such as computer learning and processing of existing crack images. Image recognition technology based on neural network is an important field of artificial intelligence. This technology can provide users with a lot of convenience. For example, the identification of plant disease pictures can quickly find diseased plants; Accurately obtain identity information. However, due to the unique narrow width of cracks and the difficulty of marking small cracks, the existing neural network identification methods are poor in crack identification, and the accuracy of identification is low. At present, neural networks are more suitable for ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/00G06T5/00G06T5/20G06K9/46G06K9/62G06T17/00
CPCG06T7/001G06T5/002G06T5/20G06T17/00G06T2207/20081G06T2207/20084G06T2207/30132G06V10/462G06F18/22
Inventor 章杨松顾天纵张宁何元李孟寒
Owner NANJING UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap